Assigning officers to high crime areas reduces crime, who knew...

February 11, 2015 8:30 AM

Forbes brings us the case of multiple police departments across the nation that use PredPol (Predictive Policing) to help figure out where crimes are likely to occur in the future based on what crime occured at what time(when) and in what location (where).

The company's claim to fame is that by using computers and "The Cloud" they can predict where crimes are more likely to occur. Here I thought the police department had this information (in the raw crime reports) and just needs some data massaging and statistical analysis to see that areas that have crime at certain parts of the day are more likely than others to have recurrences of crime as per the Broken windows therory.

posted by Hasteur (25 comments total)

Looks like we're safe until they figure out time-travel, in which case, we've seen all the movies on how that works out.
posted by k5.user at 8:31 AM on February 11, 2015 [2 favorites]

(and lack of real analysis hurts .. Given the broken window theory has been discounted greatly given crime rates have dropped across the board, regardless of whether town/city have implemented such policies or not .. )
posted by k5.user at 8:34 AM on February 11, 2015 [3 favorites]

PredPol (Predictive Policing) to help figure out where crimes are likely to occur in the future... by using computers and "The Cloud" they can predict where crimes are more likely to occur.

By "Cloud," they mean "three bald people in a pool," right?
posted by entropicamericana at 8:35 AM on February 11, 2015 [26 favorites]

We lasted a whole five minutes before making that joke. Impressive!
posted by jbickers at 8:43 AM on February 11, 2015 [2 favorites]

THE NERVE OF PEOPLE MAKING JOKES ON METAFILTER
posted by entropicamericana at 8:46 AM on February 11, 2015 [21 favorites]

There isn't really that much relationship between PredPol and the broken windows theory. Broken windows theory is like a "miasma" theory of disease transmission where PredPol is more like a germ theory. That is, "broken windows" says that certain kinds of low level lawlessness foster a general disregard for the laws and therefore prescribes a constant effort to crack down on minor misdemeanors, while PredPol is focused on significant criminal activities that are likely--not because of their "atmospheric" pervasiveness but because of their particular nature--to generate subsequent criminal acts (by "contagion" if you will).
posted by yoink at 8:54 AM on February 11, 2015 [11 favorites]

This sounds more like a commercialized version of hotspots policing, which is supported by research, than broken windows policing. I'm not sure why departments need special software to do this, they generate the data that the program would use in the first place.

Fun combination of quotes though:

"Others worry that police chiefs and city governments will rush into the embrace of Big Data without understanding how it works."

and

“I’m not really concerned about the formulas,” said Atlanta Police Chief George Turner, who implemented the software in July 2013. “That’s not my business. My business is to fight crime in my city.”
posted by _cave at 9:12 AM on February 11, 2015 [4 favorites]

I'm not sure why departments need special software to do this, they generate the data that the program would use in the first place.

Having data isn't the same as analyzing it, though.

“That’s not my business. My business is to fight crime in my city.”

Woo, awesome. Who cares where the impetus to police this area comes from, just do it!
posted by Steely-eyed Missile Man at 9:18 AM on February 11, 2015 [1 favorite]

By "Cloud," they mean "three bald people in a pool," right?

I DON'T GET IT!
posted by Mental Wimp at 9:33 AM on February 11, 2015 [2 favorites]

I DON'T GET IT!

You will!
posted by yoink at 9:35 AM on February 11, 2015 [10 favorites]

It sounds like Hyde Park in Chicago, where I lived as a U of C student, is a pretty good example of why this works and why it doesn't work. The crime rate was orders of magnitude lower than the surrounding neighborhoods because UCPD officers were patrolling constantly and no kidding, people don't like to commit crimes in front of cops. It also made a lot of people pretty upset, understandably, because they couldn't go about their business without feeling like they were under constant surveillance for being poor and black.
posted by capricorn at 9:38 AM on February 11, 2015 [4 favorites]

I wonder if the actual effect exact is the opposite of "broken windows" policing. Rather than discourage criminals with a display of order, it perturbs criminals by disturbing their sense of order when they see police in unexpected places. I wonder if the actual locations matter, as long as their distribution is implausible, difficult to predict, but somewhere in the general area near where a crime has been committed recently.

That might explain why the police stopped using the system in Louisiana; being inexplicably preceded by an invisible wave of crime interdiction for long enough might be demoralizing. Especially if it means you have to park your car in some bullshit place on a shitty street that you would never park on, but the robot told you to do it so here we are.

Aside, is Forbes still distributing viruses in its interstitial?
posted by ethansr at 9:43 AM on February 11, 2015 [1 favorite]

"Others worry that police chiefs and city governments will rush into the embrace of Big Data without understanding how it works."

I'm worried that the author doesn't understand how "Big Data" works. Unless you're dealing with hundreds of terabytes, it's just plain old data. Of course, taking away the buzzword du jour makes this just plain old statistics and statistics is yucky and who wants to see a puppy?
posted by indubitable at 9:53 AM on February 11, 2015 [6 favorites]

By "Cloud," they mean "three bald people in a pool," right?

I DON'T GET IT!

Here you go. Have fun watching Tom and his snazzy screens and pooled bald people interact!
posted by bearwife at 9:57 AM on February 11, 2015

It sounds like Hyde Park in Chicago, where I lived as a U of C student, is a pretty good example of why this works and why it doesn't work. The crime rate was orders of magnitude lower than the surrounding neighborhoods because UCPD officers were patrolling constantly and no kidding, people don't like to commit crimes in front of cops. It also made a lot of people pretty upset, understandably, because they couldn't go about their business without feeling like they were under constant surveillance for being poor and black.

But Hyde Park is not a crime hot spot. It is a Forward Operating Base of a private army (UCPD are a weird unaccountable private police force that has been granted state level authority because shenanigans) which has been tasked with the implicit mission of making the locals uncomfortable in the stolen portion of their own neighborhood. UCPD are not merely watchful. They are aggressively discouraging certain residents from entering a secured enclave. The hostile police presence of that place creeps me out and I am not even the target of their hostility.

There is a big difference between targetted policing and Hyde Park's defacto border patrol.
posted by srboisvert at 10:12 AM on February 11, 2015 [3 favorites]

while PredPol is focused on significant criminal activities that are likely--not because of their "atmospheric" pervasiveness but because of their particular nature--to generate subsequent criminal acts (by "contagion" if you will).

So instead of a wooden ball, it's more like a regional hue alert
posted by pwnguin at 10:17 AM on February 11, 2015

stolen portion of their own neighborhood

This is mostly unintelligible. First, because it makes some kind of sin out of people choosing a place to live. Second, because Hyde Park has been primarily a community for the relatively affluent from the beginning. Even the University of Chicago, the main reason that a middle-class still clings to this part of Chicago, considerably predates the Great Migration. Misguided hatred of migration and change would have to be directed in the other direction.
posted by Winnemac at 11:30 AM on February 11, 2015 [1 favorite]

'Predictive Policing' Company Uses Bad Stats, Contractually-Obligated Shills To Tout Unproven 'Successes' (Techdirt)
posted by RobotVoodooPower at 11:57 AM on February 11, 2015 [1 favorite]

This is mostly unintelligible. First, because it makes some kind of sin out of people choosing a place to live. Second, because Hyde Park has been primarily a community for the relatively affluent from the beginning. Even the University of Chicago, the main reason that a middle-class still clings to this part of Chicago, considerably predates the Great Migration. Misguided hatred of migration and change would have to be directed in the other direction.
I don't disagree that I was a bit unintelligible but you seem to be willfully obtuse as well.

It wasn't people choosing where to live when eminent domain seizures, redlining and apartheid style covenants segregated thse communities in Chicago's post-war era. It was white people choosing they could live and also where black people could and couldn't live.

The University of Chicago used eminent domain to seize property that black people were already living in or property they even feared black people might possibly live in order to raze it and drive them out. This is described, very generously on some University of Chicago web pages as "displacing some people" and "raising the socio-economic profile"

Hyde Park as it currently exists is the outcome of systematic and deliberate structural racism. Hyde Park's particular and very peculiar current police situation is an active enforcement vestige of that. It isn't hot spot policing. It's border control.
posted by srboisvert at 12:48 PM on February 11, 2015 [3 favorites]

Here you go. Have fun watching Tom and his snazzy screens and pooled bald people interact!

Duh, Wimp, duh.
posted by Mental Wimp at 2:28 PM on February 11, 2015

As in quantum physics, as in psychological studies, as in political polls: if what you're measuring might be affected by the consequences of measuring it, you'd better be methodical to the Nth power or you're going to have a Bad Time.

Or more likely, you're going to ensure that others have a Bad Time compensating for your screwups.
posted by Riki tiki at 7:23 PM on February 11, 2015

Ferguson Is Being Sued Over ‘Debtors Prisons’ For the Poor
posted by jeffburdges at 7:18 AM on February 12, 2015

Cops Have Killed Every 8 Hours in 2015, Sending At Least Three People to Early Graves Per Day
posted by jeffburdges at 11:34 AM on February 18, 2015

New York is illegally shackling pregnant incarcerated women

Did the US prison boom lead to the crime drop? New study says no
posted by jeffburdges at 6:02 AM on February 23, 2015

Theater of Justice by Molly Crabapple on Cecily McMillan
posted by jeffburdges at 6:04 AM on February 23, 2015